Paper Title
REMOVAL AND IMPROVEMENT OF RAIN STREAKS OR SNOWFLAKES ON IMAGES USING A MATCHED FILTERAbstract
On this paper, we suggest the green rules for removing rain or snow from a single color photograph. Our references take advantage of the popular techniques used in photographic processing, in particular, image decay and dictionary study. Initially, a mixture of rain / snow detection and guided smooth out is used to decompose the input image into the perfect complementary pair: (1) the low-frequency element with no rain or snow is almost true and (2) the current rain / snow problem is not spectacular, and some or all of the photo Contains a lot of data. Then, we have a passion for capturing photographic information from high-frequency details. To save you this, we lay out a 3-layer hierarchical scheme. In the number one layer, there is an over-the-top dictionary skill and three classifications are completed to classify the unpredictable-frequency frequency problem as rain / snow and non-rain / snow additions, in which some general trends of rain / snow are applied. 2D Within one layer, each other combination of rain / snow detection and guided filtering is completed at the rain / snow details obtained within the first layer. In 1/3 layer, the sensitivity of the variation during shadow channels (SVCC) is calculated to decorate the visible high-quality of the rain / snow-removed film. The effectiveness of our set of guidelines will be examined through each subjective (seemingly great) and objective (by providing rain / snow on some ground-reality pics) strategies that indicate dominance over most modern-modern works.
KEYWORDS : Rain and snow removal, picture decomposition, dictionary reading, guided filtering, sparse example.